A method, apparatus, equipment and medium for correcting CBCT image truncation artifacts

Abstract

This application provides a method, apparatus, device, and medium for correcting truncation artifacts in CBCT images. The method includes: scanning a target region using a dual-energy CBCT device to acquire high-energy projection data and low-energy projection data of the target region; performing matrix material decomposition processing on the high-energy projection data and low-energy projection data to obtain the soft tissue projection and bone projection of the target region; locating the truncated region of the original CBCT image to be corrected, and performing edge fitting on the truncated region based on the edge features of the soft tissue projection to generate an edge fitting curve for the truncated region; using the edge fitting curve as the boundary and combining the grayscale information of the soft tissue projection, completing the truncated region and fusing it with the non-truncated region of the original CBCT image to obtain the corrected CBCT image. This application can effectively eliminate truncation artifacts, improve the integrity and accuracy of CBCT images, and meet the clinical diagnostic needs for observing edge details.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a method, apparatus, device and medium for correcting CBCT image truncation artifacts. Background Technology

[0002] In clinical applications of CBCT (Cone Beam Computed Tomography) medical imaging technology, such as in head disease examinations and oral diagnosis, the field of view (FOA) for image reconstruction is determined by the physical size of the detector. Once the effective detection area of ​​the detector is fixed, the boundary of the FOA is also determined, and tissue image information outside this range cannot be effectively captured. Due to the physical bottleneck of detector size, the edges of the CBCT images obtained from actual scans often appear as follows: Figure 1 The "truncated image" shown refers to the direct truncation of tissue images beyond the field of view (FOA), resulting in significant information loss or artifacts in the marginal areas (especially the junction of soft tissue and bone in areas such as the head). This type of problem not only disrupts the integrity of the image, but if the truncated area happens to cover a potential lesion, it may directly interfere with the doctor's judgment of the condition; even if the truncated area does not involve a lesion, the incomplete marginal information will still reduce the overall image quality, affecting the reliability and accuracy of clinical diagnosis and posing risks to subsequent treatment decisions.

[0003] To address the aforementioned truncated shadow problem, related technologies primarily employ two processing schemes: one is based on directly fitting and completing the edges of the truncated region using the original monoenergetic projection. However, soft tissue and bone tissue coexist in target areas such as the head, and the density difference between these two types of tissue is significant, resulting in large differences in their X-ray absorption coefficients. Traditional fitting methods cannot specifically eliminate the interference of bone tissue on edge grayscale, leading to residual artifacts or grayscale information deviations in the completed edges, making it difficult to meet the image accuracy requirements for clinical diagnosis. The other approach relies on iterative reconstruction algorithms for optimization. While these algorithms are effective in handling scenarios with limited viewing angles (such as insufficient scanning angles due to partial occlusion), they offer no significant advantage in addressing the truncated shadow completion problem caused by limited projection field of view. This is because the core of this algorithm lies in iteratively optimizing the projection and backprojection processes to reduce reconstruction errors. However, the essence of truncated shadows is data loss, not simply insufficient viewing angle coverage. Therefore, during the iterative process, it is not only difficult to effectively fill in the true tissue information of the truncated region, but it also continuously propagates and amplifies the errors caused by the initial truncated data, resulting in the truncated shadow problem remaining prominent in the final reconstructed image, making accurate completion impossible. Summary of the Invention

[0004] To overcome the problems existing in related technologies, this application provides a method, apparatus, device and medium for correcting CBCT image truncation artifacts.

[0005] According to a first aspect of the embodiments of this application, a method for correcting truncation artifacts in CBCT images is provided, the method comprising:

[0006] The target area is scanned using a dual-energy CBCT device to obtain high-energy projection data and low-energy projection data of the target area;

[0007] The high-energy projection data and low-energy projection data are subjected to matrix material decomposition processing to obtain the soft tissue projection and bone projection of the target region.

[0008] Locate the truncated region of the original CBCT image to be corrected, and perform edge fitting on the truncated region based on the edge features of the soft tissue projection to generate the edge fitting curve of the truncated region.

[0009] Using the edge fitting curve as the boundary and combining the grayscale information of the soft tissue projection, the truncated region is completed and fused with the non-truncated region of the original CBCT image to obtain the corrected CBCT image.

[0010] According to a second aspect of the embodiments of this application, a device for correcting truncation artifacts in CBCT images is provided, the device comprising:

[0011] The dual-energy projection data acquisition module is used to scan a target area using a dual-energy CBCT device to acquire high-energy projection data and low-energy projection data of the target area.

[0012] The base material projection decomposition module is used to perform base material decomposition processing on the high-energy projection data and low-energy projection data to obtain the soft tissue projection and bone projection of the target area.

[0013] The edge curve fitting module is used to locate the truncated region of the original CBCT image to be corrected, and to perform edge fitting on the truncated region based on the edge features of the soft tissue projection, thereby generating the edge fitting curve of the truncated region.

[0014] The image correction module is used to complete the truncated region by using the edge fitting curve as the boundary and combining the grayscale information of the soft tissue projection, and then merge it with the non-truncated region of the original CBCT image to obtain the corrected CBCT image.

[0015] According to a third aspect of the embodiments of this application, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described in the first aspect.

[0016] According to a fourth aspect of the embodiments of this application, a computer-readable storage medium is provided, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the method described in the first aspect.

[0017] The technical solutions provided in this application embodiment may include the following beneficial effects:

[0018] In this embodiment, high- and low-energy projection data of the target area are acquired by a dual-energy CBCT device. The soft tissue projection without bone tissue interference is obtained by decomposing the matrix material. The edge fitting curve of the truncated area of ​​the original single-energy CBCT image is then optimized by the edge features of the soft tissue projection. The truncated area is then completed by combining the grayscale information of the soft tissue projection, thereby effectively eliminating truncation artifacts, improving the integrity and accuracy of CBCT images, and meeting the clinical diagnostic needs for observing edge details.

[0019] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0020] The accompanying drawings, which are incorporated in and form part of this application, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0021] Figure 1 This is a schematic diagram of truncation artifacts in a CBCT image according to an exemplary embodiment of this application.

[0022] Figure 2 This is a flowchart illustrating a method for correcting truncation artifacts in CBCT images according to an exemplary embodiment of this application.

[0023] Figure 3 This is a schematic projection of a target region before dual-energy decomposition according to an exemplary embodiment of this application.

[0024] Figure 4 This application illustrates a dual-energy decomposition soft tissue projection map and a bone projection map according to an exemplary embodiment.

[0025] Figure 5a This is a schematic diagram of a CBCT image before correction according to an exemplary embodiment of this application.

[0026] Figure 5b This is a schematic diagram of a CBCT image after correction according to an exemplary embodiment of this application.

[0027] Figure 6 This is a schematic diagram of a truncation artifact correction device for CBCT images according to an exemplary embodiment of this application.

[0028] Figure 7 This is a schematic diagram of the structure of a computer device according to an exemplary embodiment of this application. Detailed Implementation

[0029] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0030] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0031] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0032] In clinical applications of CBCT medical imaging technology, such as head disease examinations and oral diagnosis, the field of view (FOA) of image reconstruction is determined by the physical size of the detector. Once the effective detection area of ​​the detector is fixed, the boundary of the FOA is also determined, and tissue image information outside this range cannot be effectively captured. Due to the physical bottleneck of detector size, the edges of the CBCT images obtained from actual scans often appear as follows: Figure 1 The "truncated area" indicated by the arrow refers to the direct truncation of tissue images beyond the field of view (FOA), resulting in significant information loss or artifacts in the marginal areas (especially the junction of soft tissue and bone in areas such as the head). This type of problem not only disrupts the integrity of the image, but if the truncated area happens to cover a potential lesion, it may directly interfere with the doctor's judgment of the condition; even if the truncated area does not involve a lesion, the incomplete marginal information will still reduce the overall image quality, affecting the reliability and accuracy of clinical diagnosis and posing risks to subsequent treatment decisions.

[0033] To address the aforementioned truncated shadow problem, related technologies primarily employ two processing schemes: one is based on directly fitting and completing the edges of the truncated region using the original monoenergetic projection. However, soft tissue and bone tissue coexist in target areas such as the head, and the density difference between these two types of tissue is significant, resulting in large differences in their X-ray absorption coefficients. Traditional fitting methods cannot specifically eliminate the interference of bone tissue on edge grayscale, leading to residual artifacts or grayscale information deviations in the completed edges, making it difficult to meet the image accuracy requirements for clinical diagnosis. The other approach relies on iterative reconstruction algorithms for optimization. While these algorithms are effective in handling scenarios with limited viewing angles (such as insufficient scanning angles due to partial occlusion), they offer no significant advantage in addressing the truncated shadow completion problem caused by limited projection field of view. This is because the core of this algorithm lies in iteratively optimizing the projection and backprojection processes to reduce reconstruction errors. However, the essence of truncated shadows is data loss, not simply insufficient viewing angle coverage. Therefore, during the iterative process, it is not only difficult to effectively fill in the true tissue information of the truncated region, but it also continuously propagates and amplifies the errors caused by the initial truncated data, resulting in the truncated shadow problem remaining prominent in the final reconstructed image, making accurate completion impossible.

[0034] Based on this, to address the problems of poor truncation artifact completion and residual edge artifacts in CBCT images in related technologies, this application provides a method for truncation artifact correction in CBCT images. This method acquires high- and low-energy projection data of the target region using a dual-energy CBCT device, obtains soft tissue projection without bone tissue interference through matrix material decomposition, optimizes the edge fitting curve of the truncation region in the original single-energy CBCT image using the edge features of the soft tissue projection, and completes the truncation region by combining the grayscale information of the soft tissue projection. This effectively eliminates truncation artifacts, improves the integrity and accuracy of CBCT images, and meets the clinical diagnostic needs for observing edge details.

[0035] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0036] Figure 2 This is a schematic flowchart illustrating a method for correcting truncation artifacts in CBCT images according to an exemplary embodiment of this application. Figure 2 As shown, the method includes steps S201 to S204.

[0037] Step S201: Scan the target area using a dual-energy CBCT device to obtain high-energy projection data and low-energy projection data of the target area.

[0038] In this embodiment, the target area can be a site prone to truncation artifacts in clinical practice (such as the head, oral cavity, etc.). This area contains both soft tissue (scalp, muscles, brain tissue, etc.) and bone tissue (skull, jawbone, etc.), and the two types of tissue have significantly different X-ray attenuation. Traditional single-energy completion methods are easily affected by bone tissue interference. Therefore, dual-energy projection data acquisition is necessary to lay the foundation for subsequent accurate correction. During the scanning process, the dual-energy CBCT device can achieve high-energy and low-energy scanning by switching the tube voltage or tube current of the X-ray tube. For example, a tube voltage parameter of 110kVp can be used for high-energy scanning, and a tube voltage parameter of 80kVp can be used for low-energy scanning. During scanning, the device can rotate around the target area to acquire images. For example, the rotation angle range can be set to 360°, and a frame of two-dimensional projection image can be acquired at fixed angle intervals. Then, two-dimensional projection data sequences at corresponding energies are acquired as high-energy projection data and low-energy projection data of the target area, providing a data basis for subsequent soft tissue and bone tissue projection decomposition.

[0039] Step S202: Perform matrix material decomposition processing on the high-energy projection data and low-energy projection data to obtain the soft tissue projection and bone projection of the target area.

[0040] In this embodiment, the core of the base material decomposition is based on the X-ray attenuation law and the base material equivalence principle, which is the key theoretical basis for achieving projection separation of soft tissue and bone tissue. From a principle perspective, when X-rays of different energies pass through human tissue, their attenuation degree changes with variations in the tissue's atomic number, density, and thickness. The base material equivalence principle states that the attenuation effect of any complex tissue on X-rays can be equivalent to the superposition of attenuations from two "base materials." By acquiring CBCT projection data at both high and low energies, the attenuation information of the mixed tissue within the target area can be decomposed into the attenuation contributions of the two base materials, thereby separating the independent projection information of specific tissues (such as soft tissue and bone tissue), providing an interference-free tissue imaging basis for subsequent edge optimization of the truncated region.

[0041] Specifically, looking at the formula derivation, according to the Lambert-Beer law, the intensity of photons received by the detector after X-rays pass through an object... With incident intensity The relationship satisfies:

[0042]

[0043] in, This represents the path of X-rays through the object. For a point on the path Organizations in energy The linear attenuation coefficient is calculated. However, since the detector in a CBCT device actually receives an "energy spectrum integration signal" (non-monochromatic X-rays), the energy spectrum of the X-ray source needs to be introduced. Distribution (unit: number of photons / ( At this point, the actual projected grayscale value (and (Proportional) can be expressed as:

[0044]

[0045]

[0046] Based on the principle of equivalent base materials, Substituting into the attenuation formula, the path The total attenuation can be decomposed into the superposition of attenuations of the two base materials, that is:

[0047]

[0048] in, For the base material coefficient and Along the path The projected value (unit: mm) is also the target that the dual-energy decomposition needs to solve directly.

[0049] If the measured projection data (i.e., the degree of ray attenuation) are the same, then the base material thickness combination and Numerically, the projected value of the base material coefficient ( , The data show a one-to-one correspondence, which provides theoretical support for subsequently inferring the thickness of the base material from the projection data.

[0050] For diagnostic imaging, the energy of kV-level X-rays is... The attenuation of monochromatic X-rays after passing through a material can be linearly expressed by the linear attenuation coefficients of the two base materials, that is:

[0051]

[0052] in, and The two base materials are respectively in terms of energy Linear attenuation coefficient (unit: When using high energy With low energy Two sets of attenuation relationships can be obtained when scanning with two monochromatic X-rays:

[0053]

[0054]

[0055] in, , These represent the emission and incident intensities of low-energy rays, respectively. , These represent the outgoing and incoming intensities of high-energy rays, respectively.

[0056] The core task of bienergy decomposition is to project grayscale values ​​using known high and low energies. , (Corresponding to the grayscale information of pixels in high-energy and low-energy projection data respectively), solve the above nonlinear equations to obtain the projection value of the base material. , Among them, the total amount of low-energy decay Total high-energy decay The solution process requires first establishing a projection solution model for the basic material coefficients, that is, determining... , and , The functional relationship can be specifically achieved through an approximation formula:

[0057]

[0058]

[0059] The values ​​of parameters a0, a1, a2, a3, a4, a5, b0, b1, c0, c1, c2, c3, c4, c5, d0, and d1 in the formula can be determined by the relationship between the base material data established through preliminary experiments.

[0060] In practical terms, to efficiently complete the decomposition of the substrate material, a lookup table of the correspondence between "dual-energy projection grayscale value - substrate material thickness combination" can be pre-constructed: dual-energy projection images of soft tissue and bone tissue under different thickness combinations are collected experimentally, and the low-energy grayscale value corresponding to each thickness combination is recorded. With high energy grayscale value This data is organized into a lookup table and stored, which can be directly called during subsequent decomposition to avoid duplicate calculations.

[0061] The specific decomposition steps are as follows: First, extract the high-energy grayscale value of each pixel from the high-energy projection data obtained in step S201. Simultaneously, the low-energy grayscale values ​​of the corresponding pixels are extracted from the low-energy projection data. Here, "corresponding pixel" refers to the same coordinate point of the same scan position in high- and low-energy projection, ensuring a one-to-one correspondence between grayscale values ​​and actual tissue locations; then, in a pre-built lookup table, the pixel is used as the reference point. and Use the search criteria to find matching soft tissue thicknesses. (i.e., the base material projection value in the above formula) ) and bone thickness (i.e., base material projection value) Finally, the soft tissue thickness of all pixels is fitted according to their coordinate positions to generate a soft tissue projection containing only soft tissue attenuation information. Similarly, a bone projection containing only bone attenuation information is generated based on the bone thickness of all pixels. The soft tissue / bone projection images before and after dual-energy decomposition can be found in the reference images. Figure 3 , 4 This allows for a direct and clear observation of the separation effect between the two tissue projections.

[0062] It should be noted that the selection of the base material in this embodiment is not random, but follows the core principle of "equivalent coverage + energy spectrum differentiation": "equivalent coverage" requires a combination of the linear attenuation coefficients of the two base materials ( , It needs to be able to approximately cover the attenuation characteristics of all tissues within the target imaging area (such as the head), that is, the linear attenuation coefficient of any tissue within that area. All can be accessed through and The two base materials are superimposed with specific weights; "energy spectrum differentiation" requires that the attenuation difference between the two base materials be significant in the high and low energy ranges to ensure that the projection information of the two tissues does not overlap or interfere after decomposition. Considering that this application is aimed at CBCT truncation artifact correction in areas such as the head, which mainly include soft tissue (such as scalp and brain tissue) and bone (such as skull and jawbone), and that the attenuation difference of these two tissues to X-rays is significant, fully satisfying the principles of "equivalent coverage" and "energy spectrum differentiation", soft tissue and bone tissue can be selected as base materials, providing accurate tissue imaging basis for subsequent optimization of the truncation area edge.

[0063] Step S203: Locate the truncated region of the original CBCT image to be corrected, and perform edge fitting on the truncated region based on the edge features of soft tissue projection to generate the edge fitting curve of the truncated region.

[0064] In this embodiment, the original CBCT image to be corrected is the single-energy CBCT image directly obtained in clinical diagnosis. The location of its truncated region can be performed by combining the dual logic of "geometric constraints" and "grayscale mutation detection" to ensure the accuracy of the location. This avoids missing the actual truncated region due to simple geometric judgment and also eliminates false positive grayscale mutation regions caused by interference such as metal artifacts.

[0065] Specifically, CBCT equipment allows users to preset the geometric parameters of the field of view (FOV) before scanning, such as presetting the circular FOV scanning radius R and setting the image center as the isocenter of the scan. (i.e., the center reference point of the target area during scanning).

[0066] First, candidate truncation regions are selected based on geometric constraints: Calculate the truncation region for each pixel in the original CBCT image. To the scanning center The distance d can be expressed by the following formula: Considering the slight positioning deviations that may occur during the scanning process, an edge error can be reserved, such as 2 pixels, when the distance of a certain pixel is... When a pixel exceeds the effective imaging range of the field of view (FOV), it is determined that the pixel exceeds the FOV and is marked as a candidate truncation region for geometric constraints. Subsequently, another set of candidate truncation regions is screened through gray-level abrupt change detection: the core feature of the truncation region is "sudden loss of image information," which is reflected in the gray-level as a significant jump in the gray-level value of adjacent pixels. Therefore, this embodiment can also use gray-level abrupt change detection, such as the Sobel gradient algorithm, to perform gray-level abrupt change detection on the original CBCT image. By calculating the gray-level gradient value (i.e., gray-level abrupt change value) of each pixel in the horizontal and vertical directions, the gray-level difference between it and the surrounding pixels is quantified. A preset gray-level abrupt change threshold is set (this threshold is determined based on the common gray-level range of the CBCT image of the target area). When the gray-level gradient value of a certain pixel exceeds the threshold, it is determined that it is in a gray-level abrupt change region and is marked as a candidate truncation region for gray-level abrupt change. Finally, to further improve the accuracy of truncation region localization, the intersection of the geometrically constrained candidate truncation region and the gray-scale abrupt change candidate truncation region is taken as the final truncation region: geometric constraints ensure that the region conforms to the physical limitations of FOV in space, and gray-scale abrupt change detection ensures that the region conforms to the appearance of the truncation shadow in image features. The combination of the two can effectively eliminate false positive gray-scale abrupt change regions caused by metal implants, etc. (although such regions have gray-scale jumps, they do not exceed the geometric range of FOV and do not belong to the truncation region), while avoiding geometric misjudgment regions caused by small deviations in FOV parameters, and finally obtain the accurate range of the truncation region.

[0067] After the truncated region is located, the edge of the truncated region can be fitted based on the soft tissue projection obtained in step S202 to generate an edge fitting curve.

[0068] Traditional methods directly fit the edge points of the truncated region in the original monoenergy CBCT image (e.g., linear fitting, quadratic polynomial fitting). However, the edges of the truncated region in the original monoenergy image are affected by bone tissue interference (e.g., the skull edge overlaps with the truncated boundary), resulting in large gray-level fluctuations at the edge points and a "sawtooth" deviation in the fitting curve. In this embodiment, the soft tissue projection obtained in step S202 has eliminated bone tissue interference through bienergy decomposition (the bone tissue has been separated into separate bone projections). Its truncated region edge is composed only of soft tissue (e.g., scalp, muscle), with continuous gray-level distribution and clear boundary contours, providing a "pure edge reference". By mapping the edge features of the soft tissue projection onto the original monoenergy CBCT image, the edge fitting deviation caused by bone tissue in the monoenergy image can be corrected.

[0069] The specific fitting steps are as follows: First, perform soft tissue projection edge extraction: The same gray-level change detection method as the original CBCT image (such as Sobel gradient + threshold screening) can be used to perform gray-level change detection on the soft tissue projection image to ensure the consistency of the edge extraction logic; extract the pixels in the soft tissue projection that meet the preset gray-level change features to form the soft tissue projection truncated edge point set P, which directly reflects the real edge contour of the soft tissue in the truncated area. Next, edge point screening and piecewise curve fitting are performed: To further eliminate potentially residual noise points (such as isolated gray-level abrupt changes caused by scanning noise), the least squares method can be used to calculate the local slope of the edge point set P. For example, five adjacent edge points can be used as a calculation window to solve the gray-level gradient slope within each window. Edge points whose slope changes meet the preset continuity conditions (such as slope change amplitude being less than the preset change threshold) are selected. The slopes of these points are continuous and conform to the natural shape of soft tissue edges, so they can be used as valid edge points, and abnormal points with abrupt slope changes are eliminated (these points are mostly noise interference and not real soft tissue edges). After that, piecewise curve fitting is performed based on the slope change trend of the valid edge points to generate the edge fitting curve of the truncated region: For example, when the slope of a certain segment of valid edge points remains stable, linear fitting is used; when the slope changes slowly, quadratic polynomial fitting is used. Through piecewise processing, it can be ensured that the edge curve conforms to the natural contour of soft tissue and has both local continuity and overall smoothness. Finally, the fitted edge curves can be smoothed and optimized: for example, Gaussian filtering can be applied to the edge curves obtained by piecewise fitting, and the standard deviation of the Gaussian filter can be set (e.g., set to...). =0.5 (This parameter can effectively eliminate curve fluctuations caused by minor noise while preserving the edge contour). By filtering, the curve details are further smoothed and residual noise is eliminated. Finally, the optimized edge fitting curve of the truncated area is obtained. This curve can accurately reflect the real edge morphology of soft tissue in the truncated area and provide a precise boundary benchmark for subsequent truncated shadow completion.

[0070] Step S204: Using the edge fitting curve as the boundary and combining the grayscale information of the soft tissue projection, the truncated region is completed and fused with the non-truncated region of the original CBCT image to obtain the corrected CBCT image.

[0071] In this embodiment, the original CBCT image can be fused and completed using a "layered processing and grayscale matching" approach. That is, the optimized edge fitting curve obtained in step S203 is used as the boundary reference. Combining the grayscale distribution pattern of the original CBCT image with the real tissue grayscale characteristics of the soft tissue projection, the image is divided into different regions for targeted processing. This ensures the authenticity of the tissue information in the completed region and avoids grayscale discontinuity between the completed region and the normal region, ultimately achieving seamless image fusion.

[0072] Specifically, the original CBCT image can be divided into regions first: based on the positional relationship between the pixels and the edge fitting curve in the original CBCT image, the image is divided into three regions: the first is the normal region (R1), which is the region inside the edge fitting curve and close to the preset scan center. This region is within the effective FOV imaging range, and the tissue grayscale information is complete, so no supplementation is needed and the original grayscale can be directly retained later; the second is the transition region (R2), which is the region with a preset width on both sides of the edge fitting curve (such as a range of 5 pixels on both sides of the edge fitting curve). This region serves as the "grayscale connection zone" between the normal region and the supplemented region. Since the grayscale sources of the normal region and the supplemented region are different (the former is the original scan data, and the latter is the soft tissue projection mapping data), direct stitching is prone to obvious grayscale jumps. Therefore, it is necessary to set a transition region to achieve smooth grayscale transition by grayscale fusion; the third is the supplemented region (R3), which is the truncated region outside the edge fitting curve and beyond the FOV range. This region has no effective tissue scan data and needs to be supplemented based on the grayscale information of soft tissue projection.

[0073] After the regions are divided, targeted grayscale processing is performed on the three regions respectively. The completion region (R3) and the transition region (R2) are the core of the processing, while the normal region (R1) can simply retain the original grayscale.

[0074] For the completed region (R3), it is necessary to generate a completed grayscale value that matches the original image based on the grayscale information of the soft tissue projection. The core of this is to achieve grayscale matching through "coordinate mapping + grayscale transformation": First, coordinate mapping is performed to transform any pixel within the completed region R3. Directly mapped to the same coordinates of the soft tissue projection Since the soft tissue projection has the same scanning field of view and spatial position as the original CBCT image, and the same coordinates correspond to the same physical location, the gray value of that location in the soft tissue projection can be directly obtained. Next, grayscale conversion is performed to avoid grayscale discrepancies in the completed area caused by differences in grayscale measurement standards between the soft tissue projection and the original image. Specifically, grayscale conversion can employ a mapping logic of "background grayscale alignment + edge grayscale matching," first obtaining two sets of key grayscale averages: one being the grayscale average of the background region in the original monoenergetic CBCT image. (Typically close to -1000 HU), and the mean edge gray value at the R2 boundary in the original image. Secondly, the average grayscale value of the background region (air region without any tissue) in the soft tissue projection. And the mean gray value of the soft tissue in the soft tissue projection of the transition region R2. Furthermore, based on the gray-scale mean of the soft tissue in the transition region of the original CBCT image... Compared with the average gray level of the background The difference, and the mean gray value of the soft tissue in the transition region of the soft tissue projection. Compared with the average gray level of the background The difference is used to determine the grayscale mapping relationship between the original CBCT image and the soft tissue projection. Based on this grayscale mapping relationship, the grayscale value of the corresponding position of the pixel in the soft tissue projection is determined. The grayscale value is converted to a padded grayscale that matches the original CBCT image, and this padded grayscale value is used as the target grayscale value for that pixel. That is, the grayscale value in the soft tissue projection is mapped using the following formula. Converted to a padded grayscale that matches the original image :

[0075]

[0076] The above formula uses background grayscale and Alignment is performed to ensure that the background grayscale of the filled area matches that of the original image; grayscale transitions are used to ensure the grayscale of the transition area matches that of the original image. and The proportional mapping ensures a seamless connection between the grayscale at the edge of the completed area and the transition area, completely avoiding grayscale discontinuities.

[0077] For the transition region (R2), a distance-weighted fusion method can be used to achieve a smooth grayscale transition: First, obtain any pixel within the transition region R2. Original grayscale values ​​in the original CBCT image And the completed grayscale value of the corresponding position of the pixel in the soft tissue projection, obtained based on the above grayscale mapping relationship. Then, calculate the pixel point. The vertical distance *d* to the edge-fitted curve is calculated, where *d* > 0 indicates the pixel is inside the curve (closer to the normal region R1), *d* < 0 indicates it is outside the curve (closer to the padded region R3), and *d* = 0 represents the curve itself. Then, based on the vertical distance *d*, the fusion weights of the original grayscale value and the padded grayscale value are determined. These weights can change linearly with distance to ensure a smooth transition of grayscale from R1 to R3. When *d* > 0, the original image grayscale fusion weights can be set to... The grayscale blending weights can be set to... (That is, the closer a pixel is to the normal region R1, the higher the fusion weight of the original grayscale value, so as to retain more original scan information); when d<0, the original image grayscale fusion weight can be set to The grayscale blending weights can be set to... (That is, the closer a pixel is to the padding region R3, the higher the fusion weight of the padding grayscale, to ensure a natural transition with the padding region.) The specific form of the fusion weight between the original grayscale value and the padding grayscale can be set according to the user's actual needs; this embodiment does not impose any restrictions. Finally, the original grayscale value and the padding grayscale are weighted based on the fusion weight to obtain the fused grayscale of the pixel, and this fused grayscale is used as the target grayscale of the pixel. The weighting formula can be... .

[0078] After the grayscale processing of the three regions is completed, the target grayscale of all pixels in the normal region, transition region, and padding region can be stitched together (i.e., the original grayscale of the normal region R1 and the blended grayscale of the transition region R2). The completed grayscale of region R3 This process yields a complete, corrected CBCT image. This image eliminates edge truncation artifacts while preserving the original image's tissue detail and authenticity, fully meeting the clinical diagnostic requirements for image integrity and accuracy.

[0079] Furthermore, to achieve further edge smoothing and noise suppression in CBCT images, the overall image can be optimized: for example, a 3×3 Gaussian filter can be applied to the transition region R2 (e.g., by setting...). =0.8) is used for smoothing to further eliminate subtle gray-level fluctuations after weighted fusion and ensure a more natural gray-level transition near the edge curve; and non-local mean filtering is used on the completed region R3 (such as setting a 7×7 search window and a 3×3 similarity window) to suppress scanning noise while preserving the soft tissue edge features of the completed area to the greatest extent (non-local mean filtering can reduce noise by matching the gray-level distribution of similar pixel blocks, avoiding the blurring of edges caused by traditional filtering).

[0080] To more intuitively demonstrate the effect of the proposed solution on correcting truncation artifacts in CBCT images, Figure 5a , 5b This image shows a comparison of head CBCT images before and after correction. Figure 5a The original CBCT image to be corrected. Figure 5b The image shown is a corrected CBCT image obtained using the method described in the embodiments of this application. From... Figure 5a It is evident that there are obvious truncation artifacts in the edge regions of the original CBCT images, while Figure 5bThe truncation artifacts in the corrected images have been effectively eliminated. Therefore, it is clear that the solution of this application embodiment can completely solve the problems of artifact residue and information deviation after completion using traditional methods, significantly improving the integrity and accuracy of CBCT images, ensuring that doctors can clearly observe tissue details in the edge areas, fully meeting the clinical diagnostic needs for observing image edge details, and providing reliable image support for subsequent disease assessment and treatment plan formulation.

[0081] Corresponding to the embodiments of the aforementioned methods, this application also provides a device for correcting truncation artifacts in CBCT images. Figure 6 This is a schematic diagram illustrating the structure of a truncation artifact correction device for CBCT images according to an exemplary embodiment of this application. Figure 6 As shown, the device includes:

[0082] The dual-energy projection data acquisition module 601 is used to scan the target area using a dual-energy CBCT device to acquire high-energy projection data and low-energy projection data of the target area.

[0083] The base material projection decomposition module 602 is used to perform base material decomposition processing on high-energy projection data and low-energy projection data to obtain the soft tissue projection and bone projection of the target area.

[0084] The edge curve fitting module 603 is used to locate the truncated region of the original CBCT image to be corrected, and to perform edge fitting on the truncated region based on the edge features of soft tissue projection, generating the edge fitting curve of the truncated region.

[0085] The image correction module 604 is used to complete the truncated region by using the edge fitting curve as the boundary and combining the grayscale information of the soft tissue projection, and then merge it with the non-truncated region of the original CBCT image to obtain the corrected CBCT image.

[0086] The specific implementation process of the functions and roles of each module in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0087] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0088] Corresponding to the embodiments of the foregoing methods, this application also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor; wherein, when the processor executes the computer program, it implements the steps of the truncation artifact correction method for CBCT images described in any of the above embodiments.

[0089] For example, processors include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), or field-programmable gate arrays (FPGAs).

[0090] For example, the memory may include at least one type of storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc.

[0091] Figure 7 This is a structural block diagram of a computer device according to an exemplary embodiment of this application. Figure 7 As shown, at the hardware level, the computer device includes a processor 701, an internal bus 702, a network interface 703, memory 704, and non-volatile memory 705, and may also include other hardware required for business operations. One or more embodiments of this application can be implemented in software, for example, the processor 701 reads the corresponding computer program from the non-volatile memory 705 into the memory 704 and then runs it. Of course, in addition to software implementation, one or more embodiments of this application do not exclude other implementation methods, such as logic devices or a combination of hardware and software, etc. That is to say, the execution subject of the above processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0092] Corresponding to the embodiments of the foregoing methods, this application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the truncation artifact correction method for CBCT images described in any of the above embodiments.

[0093] Corresponding to the embodiments of the foregoing methods, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the truncation artifact correction method for CBCT images described in any of the above embodiments.

[0094] The foregoing has described specific embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0095] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention filed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not claimed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the foregoing claims.

[0096] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

[0097] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for truncation artifact correction of a CBCT image, the method comprising: include: The target area is scanned using a dual-energy CBCT device to obtain high-energy projection data and low-energy projection data of the target area; The high-energy projection data and low-energy projection data are subjected to matrix material decomposition processing to obtain the soft tissue projection and bone projection of the target region. Locate the truncated region of the original CBCT image to be corrected, and perform edge fitting on the truncated region based on the edge features of the soft tissue projection to generate the edge fitting curve of the truncated region. Based on the positional relationship between the pixels in the original CBCT image and the edge fitting curve, the original CBCT image is divided into a normal region, a transition region, and a complete region; wherein, the normal region is the non-truncated region inside the edge fitting curve and close to the preset scan isocenter, the transition region is the region with a preset width on both sides of the edge fitting curve, and the complete region is the truncated region outside the edge fitting curve. For any pixel in the normal region, the original gray value of that pixel in the original CBCT image is taken as the target gray value of that pixel. For any pixel within the completed area, a target grayscale value for that pixel is generated based on the grayscale value of the corresponding position in the soft tissue projection. For any pixel within the transition region, the target grayscale of the pixel is generated by fusing the original grayscale value of the pixel in the original CBCT image and the grayscale value of the corresponding position of the pixel in the soft tissue projection. By stitching together the target grayscale values ​​of all pixels in the normal region, transition region, and complete region, the corrected CBCT image is obtained.

2. The method of claim 1, wherein, The high-energy projection data and low-energy projection data are subjected to matrix material decomposition processing to obtain the soft tissue projection and bone projection of the target region, including: Extract the high-energy grayscale value of each pixel in the high-energy projection data and the low-energy grayscale value of the corresponding pixel in the low-energy projection data respectively. In a pre-built lookup table of correspondence between dual-energy projection gray values ​​and base material thickness, the soft tissue thickness and bone thickness that match the low-energy gray values ​​and high-energy gray values ​​of each pixel are searched. The soft tissue projection is generated by fitting the soft tissue thickness of each pixel, and the bone projection is generated by fitting the bone thickness of each pixel.

3. The method of claim 1, wherein, Locate the truncated region of the original CBCT image to be corrected, including: Calculate the distance from each pixel in the original CBCT image to the preset scan isocenter, and mark the pixels whose distance exceeds the sum of the preset scan radius and the edge error as geometric constraint candidate truncation regions; The original CBCT image is subjected to gray-level abrupt change detection, and pixels with gray-level abrupt change values ​​exceeding a preset threshold are marked as gray-level abrupt change candidate truncation regions. The intersection of the geometric constraint candidate truncation region and the gray-scale abrupt change candidate truncation region is taken as the truncation region.

4. The method according to claim 1, characterized in that, Based on the edge features of the soft tissue projection, edge fitting is performed on the truncated region to generate an edge fitting curve for the truncated region, including: Perform grayscale abrupt change detection on the soft tissue projection and extract the edge point set of the truncated region in the soft tissue projection; Local slope calculations are performed on the set of edge points to filter out valid edge points whose slope changes meet preset continuity conditions; Based on the slope change trend of the effective edge points, piecewise curve fitting is performed to generate the edge fitting curve of the truncated region.

5. The method according to claim 1, characterized in that, For any pixel within the completed area, based on the grayscale value of the pixel at its corresponding position in the soft tissue projection, a target grayscale value for that pixel is generated, including: For any pixel within the completed area, extract the grayscale value of the corresponding position of that pixel in the soft tissue projection; The average background grayscale value of the original CBCT image, the average grayscale value of the soft tissue in the transition area of ​​the original CBCT image, the average background grayscale value of the soft tissue projection, and the average grayscale value of the soft tissue in the transition area of ​​the soft tissue projection are obtained respectively. Based on the difference between the mean gray value of the soft tissue in the transition region and the mean gray value of the background in the original CBCT image, and the difference between the mean gray value of the soft tissue in the transition region and the mean gray value of the background in the soft tissue projection, the gray-level mapping relationship between the original CBCT image and the soft tissue projection is determined. According to the grayscale mapping relationship, the grayscale value of the pixel at the corresponding position in the soft tissue projection is converted into a completed grayscale that matches the original CBCT image, and the completed grayscale is used as the target grayscale of the pixel.

6. The method according to claim 5, characterized in that, For any pixel within the transition region, based on the original grayscale value of that pixel in the original CBCT image and the grayscale value of the corresponding position of that pixel in the soft tissue projection, a target grayscale value for that pixel is generated by fusing the grayscale values, including: For any pixel in the transition region, obtain the original grayscale value of the pixel in the original CBCT image, and the completed grayscale value of the pixel at the corresponding position in the soft tissue projection based on the grayscale mapping relationship. Calculate the vertical distance from the pixel to the edge fitting curve; Based on the vertical distance, the fusion weight of the original grayscale value and the padded grayscale value is determined, wherein the closer the pixel is to the normal area, the higher the fusion weight of the original grayscale value, and the closer the pixel is to the padded area, the higher the fusion weight of the padded grayscale value. The original grayscale value and the padded grayscale value are weighted and calculated based on the fusion weight to obtain the fused grayscale value of the pixel, and the fused grayscale value is used as the target grayscale value of the pixel.

7. A device for correcting truncation artifacts in CBCT images, characterized in that, include: The dual-energy projection data acquisition module is used to scan a target area using a dual-energy CBCT device to acquire high-energy projection data and low-energy projection data of the target area. The base material projection decomposition module is used to perform base material decomposition processing on the high-energy projection data and low-energy projection data to obtain the soft tissue projection and bone projection of the target area. The edge curve fitting module is used to locate the truncated region of the original CBCT image to be corrected, and to perform edge fitting on the truncated region based on the edge features of the soft tissue projection, thereby generating the edge fitting curve of the truncated region. Image correction module, used for: Based on the positional relationship between the pixels in the original CBCT image and the edge fitting curve, the original CBCT image is divided into a normal region, a transition region, and a complete region; wherein, the normal region is the non-truncated region inside the edge fitting curve and close to the preset scan isocenter, the transition region is the region with a preset width on both sides of the edge fitting curve, and the complete region is the truncated region outside the edge fitting curve. For any pixel in the normal region, the original gray value of that pixel in the original CBCT image is taken as the target gray value of that pixel. For any pixel within the completed area, a target grayscale value for that pixel is generated based on the grayscale value of the corresponding position in the soft tissue projection. For any pixel within the transition region, the target grayscale of the pixel is generated by fusing the original grayscale value of the pixel in the original CBCT image and the grayscale value of the corresponding position of the pixel in the soft tissue projection. By stitching together the target grayscale values ​​of all pixels in the normal region, transition region, and complete region, the corrected CBCT image is obtained.

8. A computer device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1 to 6.

Images